@equationalapplications/core-llm-wiki 5.1.0 → 5.2.0

package/dist/{chunk-DM7BKDCV.mjs → chunk-J3N3WRK7.mjs}

@@ -64,6 +64,15 @@ var WikiBusyError = class extends Error {
     this.entityId = entityId;
   }
 };
+var WikiDuplicateHashError = class extends Error {
+  constructor(params) {
+    super(`Duplicate source hash for entity ${params.entityId}; another ref already holds this content`);
+    this.name = "WikiDuplicateHashError";
+    this.canonical = params.canonical;
+    this.sourceHash = params.sourceHash;
+    this.entityId = params.entityId;
+  }
+};
 var WikiTransactionError = class extends Error {
   constructor(message, options) {
     super(message, options);
@@ -693,6 +702,19 @@ var JobManager = class {
         return this.activeMaintenanceJobs.has(this.lockKeyFns[operation](entityId)) || this.activeMaintenanceJobs.has(this._pruneKey(entityId)) || this._isReembedActive(entityId) || this._isImportActiveFor(entityId) || this._isForgetActiveFor(entityId);
       case "prune":
         return this.activeMaintenanceJobs.has(this._pruneKey(entityId)) || this.activeMaintenanceJobs.has(this._librarianKey(entityId)) || this.activeMaintenanceJobs.has(this._healKey(entityId)) || this.activeMaintenanceJobs.has(this._ontologyBackfillKey(entityId)) || this._isReembedActive(entityId) || this._isIngestActiveFor(entityId) || this._isImportActiveFor(entityId) || this._isForgetActiveFor(entityId);
+      case "reembed":
+        return this.activeMaintenanceJobs.has(this._reembedKey(entityId)) || this.activeMaintenanceJobs.has(this._globalReembedKey()) || this.activeMaintenanceJobs.has(this._pruneKey(entityId)) || this.activeMaintenanceJobs.has(this._librarianKey(entityId)) || this.activeMaintenanceJobs.has(this._healKey(entityId)) || this.activeMaintenanceJobs.has(this._ontologyBackfillKey(entityId)) || this._isIngestActiveFor(entityId) || this._isImportActiveFor(entityId) || this._isForgetActiveFor(entityId);
+      case "global_reembed":
+        return this.activeMaintenanceJobs.has(this._globalReembedKey()) || this._isAnyMaintenanceActiveWithSuffix(":reembed") || this._isAnyMaintenanceActiveWithSuffix(":prune") || this._isAnyMaintenanceActiveWithSuffix(":librarian") || this._isAnyMaintenanceActiveWithSuffix(":heal") || this._isAnyMaintenanceActiveWithSuffix(":ontologyBackfill") || this.activeIngestJobs.size > 0 || this._isAnyMaintenanceActiveWithSuffix(":import") || this._isAnyMaintenanceActiveWithSuffix(":forget");
+      case "import":
+      case "forget": {
+        const selfKey = operation === "import" ? this._importKey(entityId) : this._forgetKey(entityId);
+        return this.activeMaintenanceJobs.has(selfKey) || this.activeMaintenanceJobs.has(this._librarianKey(entityId)) || this.activeMaintenanceJobs.has(this._healKey(entityId)) || this.activeMaintenanceJobs.has(this._ontologyBackfillKey(entityId)) || this.activeMaintenanceJobs.has(this._pruneKey(entityId)) || this._isReembedActive(entityId) || this._isIngestActiveFor(entityId) || this._isImportActiveFor(entityId) || this._isForgetActiveFor(entityId);
+      }
+      case "global_import":
+        return this.activeMaintenanceJobs.has(this._globalImportKey());
+      case "ingest":
+        return this._hasIngestJob(entityId) || this.activeMaintenanceJobs.has(this._pruneKey(entityId)) || this._isReembedActive(entityId) || this._isImportActiveFor(entityId) || this._isForgetActiveFor(entityId);
       default:
         return false;
     }
@@ -1159,11 +1181,23 @@ var IngestionService = class {
     this.ontologyService = ontologyService;
     this.promptService = promptService ?? new PromptService(this.options.config?.prompts);
   }
-  async ingestDocument(entityId, params) {
+  async ingestDocument(entityId, params, opts) {
     const sourceRef = normalizeSourceRef(params.sourceRef);
     if (!sourceRef) throw new Error("Invalid sourceRef");
     const sourceHash = normalizeSourceHash(params.sourceHash);
     if (!sourceHash) throw new Error("Invalid sourceHash (must be 64-char hex string)");
+    const onDuplicateHash = opts?.onDuplicateHash ?? "ingest";
+    if (onDuplicateHash !== "ingest") {
+      const refs = await this.entryRepo.findSourceRefsByHash(entityId, sourceHash);
+      const others = refs.filter((r) => r !== sourceRef);
+      if (others.length > 0) {
+        const canonical = others[0];
+        if (onDuplicateHash === "throw") {
+          throw new WikiDuplicateHashError({ canonical, sourceHash, entityId });
+        }
+        return { truncated: false, chunks: 0, duplicateOf: canonical };
+      }
+    }
     const maxChunkLength = params.maxChunkLength ?? this.options.config?.maxChunkLength ?? DEFAULT_MAX_CHUNK_LENGTH;
     const rawOverlap = params.chunkOverlap ?? this.options.config?.chunkOverlap ?? DEFAULT_CHUNK_OVERLAP;
     const chunkOverlap = Math.min(
@@ -1739,10 +1773,13 @@ var MaintenanceService = class {
       this.jobManager.releaseLock(op, entityId ?? "*");
     }
   }
-  async forget(entityId, params) {
+  async forget(entityId, params, opts) {
     if (params.clearAll && (params.entryId !== void 0 || params.taskId !== void 0 || params.sourceRef !== void 0 || params.sourceHash !== void 0)) {
       throw new Error("forget() clearAll is mutually exclusive with entryId, taskId, sourceRef, and sourceHash");
     }
+    if (opts?.dryRun === true) {
+      return this.forgetDryRun(entityId, params);
+    }
     this.jobManager.acquireLock("forget", entityId);
     try {
       const now = Date.now();
@@ -1803,11 +1840,42 @@ var MaintenanceService = class {
           throw new Error(`forget(${entityId}/${factId}) failed: ANN cleanup hook rejected`, { cause: this._sanitizeRankerError(hookErr) });
         }
       }
-      return { deleted: { entries: deletedEntries, tasks: deletedTasks } };
+      return params.clearAll ? { deleted: { entries: deletedEntries, tasks: deletedTasks }, metadataReset: true } : { deleted: { entries: deletedEntries, tasks: deletedTasks } };
     } finally {
       this.jobManager.releaseLock("forget", entityId);
     }
   }
+  /**
+   * Read-only dry-run preview of {@link forget}. Returns the same shape as a
+   * real call would, without acquiring the lock, opening a transaction, staging
+   * outbox events, or firing embedding hooks. Counts may be off-by-N during a
+   * concurrent real forget — this is accepted, not considered a bug.
+   *
+   * `metadataReset` mirrors the real call: only `clearAll: true` returns true,
+   * because only `clearAll` actually resets the metadata checkpoint.
+   * `standard` (sourceRef/sourceHash) returns no metadataReset field, just like
+   * the real call. `entryId`/`taskId` selectors are rejected with a clear error
+   * (the spec keeps dry-run narrow).
+   */
+  async forgetDryRun(entityId, params) {
+    if (params.entryId !== void 0 || params.taskId !== void 0) {
+      throw new Error("forget({ dryRun: true }) does not support entryId/taskId selectors; use sourceRef/sourceHash or clearAll");
+    }
+    if (params.clearAll) {
+      const entries2 = await this.entryRepo.countLiveByEntityId(entityId);
+      const tasks = await this.taskRepo.countLiveByEntityId(entityId);
+      return { deleted: { entries: entries2, tasks }, metadataReset: true };
+    }
+    const sourceRef = params.sourceRef !== void 0 ? normalizeSourceRef(params.sourceRef) : null;
+    if (params.sourceRef !== void 0 && !sourceRef) throw new Error("Invalid sourceRef");
+    const sourceHash = params.sourceHash !== void 0 ? normalizeSourceHash(params.sourceHash) : null;
+    if (params.sourceHash !== void 0 && !sourceHash) throw new Error("Invalid sourceHash (must be 64-char hex string)");
+    if (sourceRef === null && sourceHash === null) {
+      return { deleted: { entries: 0, tasks: 0 } };
+    }
+    const entries = await this.entryRepo.countLiveBySource(entityId, sourceRef, sourceHash);
+    return { deleted: { entries, tasks: 0 } };
+  }
   /** Core librarian pass (locks handled by {@link runLibrarian}). Package-internal orchestration hook. */
   async doRunLibrarian(entityId, promptOverride) {
     const events = await this.eventRepo.getRecent(entityId, 50);
@@ -2037,10 +2105,10 @@ var MaintenanceService = class {
     const validNewFacts = newFacts.map(validateFact).filter((f) => f !== null);
     const insertedFacts = [];
     const uniqueDeletedFactIds = Array.from(new Set(safeDeleted));
-    const healFactsForDedupe = (await this.entryRepo.findInferredTitlesByEntityId(entityId)).filter((f) => !safeDeletedSet.has(f.id));
     await this.db.withTransactionAsync(async (tx) => {
       await this.entryRepo.downgradeByIds(safeDowngraded, entityId, tx);
       await this.entryRepo.softDeleteByIds(safeDeleted, entityId, tx);
+      const healFactsForDedupe = await this.entryRepo.findInferredTitlesByEntityId(entityId, tx);
       for (const fact of validNewFacts) {
         const newTokens = titleTokens(fact.title);
         let skip = false;
@@ -2531,6 +2599,15 @@ var ImportExportService = class {
         const safeTitle = clip(String(fact.title ?? ""), IMPORT_TITLE_MAX);
         const safeBody = clip(String(fact.body ?? ""), IMPORT_BODY_MAX);
         clippedTextByFactId.set(fact.id, { title: safeTitle, body: safeBody });
+        let normalizedSourceRef = null;
+        if (fact.source_ref !== null && fact.source_ref !== void 0) {
+          normalizedSourceRef = normalizeSourceRef(fact.source_ref);
+          if (normalizedSourceRef === null) {
+            throw new Error(
+              `importDump: invalid source_ref ${JSON.stringify(fact.source_ref)} for entity "${entityId}" fact "${fact.id}" (must normalize to a non-empty string; see ingestDocument's sourceRef validation)`
+            );
+          }
+        }
         const factObj = {
           id: fact.id,
           entity_id: entityId,
@@ -2540,7 +2617,7 @@ var ImportExportService = class {
           confidence: fact.confidence,
           source_type: sourceType,
           source_hash: fact.source_hash,
-          source_ref: fact.source_ref,
+          source_ref: normalizedSourceRef,
           created_at: fact.created_at,
           updated_at: safeUpdatedAt,
           last_accessed_at: fact.last_accessed_at,
@@ -3553,6 +3630,6 @@ var WriteService = class {
   }
 };
 
-export { BaseRepository, DEFAULT_CHUNK_OVERLAP, DEFAULT_MAX_CHUNK_LENGTH, EmbeddingService, HEAL_BATCH_SIZE, HEAL_RECHECK_MS, HOOK_TIMEOUT_MARKER, ImportExportService, IngestionService, JobManager, MaintenanceService, MetadataRepository, ONTOLOGY_BACKFILL_BATCH_SIZE, ONTOLOGY_BACKFILL_MAX_PROMPT_CHARS, ONTOLOGY_BACKFILL_RECHECK_MS, ONTOLOGY_BACKFILL_SYSTEM_PROMPT, PromptService, PrunePartialFailureError, RetrievalService, SearchService, WikiBusyError, WikiTransactionError, WriteService, __privateAdd, __privateGet, __privateSet, chunkText, configureRandomSource, emptyManifest, entitySummaryMetaKey, extractSqliteCode, generateId, normalizeSourceHash, normalizeSourceRef, normalizeTitleKey, parseEmbedding, resolveEdgeDefinitions, resolveNodeType, safeSlice, validateInlineEdges, validateManifest };
-//# sourceMappingURL=chunk-DM7BKDCV.mjs.map
-//# sourceMappingURL=chunk-DM7BKDCV.mjs.map
+export { BaseRepository, DEFAULT_CHUNK_OVERLAP, DEFAULT_MAX_CHUNK_LENGTH, EmbeddingService, HEAL_BATCH_SIZE, HEAL_RECHECK_MS, HOOK_TIMEOUT_MARKER, ImportExportService, IngestionService, JobManager, MaintenanceService, MetadataRepository, ONTOLOGY_BACKFILL_BATCH_SIZE, ONTOLOGY_BACKFILL_MAX_PROMPT_CHARS, ONTOLOGY_BACKFILL_RECHECK_MS, ONTOLOGY_BACKFILL_SYSTEM_PROMPT, PromptService, PrunePartialFailureError, RetrievalService, SearchService, WikiBusyError, WikiDuplicateHashError, WikiTransactionError, WriteService, __privateAdd, __privateGet, __privateSet, chunkText, configureRandomSource, emptyManifest, entitySummaryMetaKey, extractSqliteCode, generateId, normalizeSourceHash, normalizeSourceRef, normalizeTitleKey, parseEmbedding, resolveEdgeDefinitions, resolveNodeType, safeSlice, validateInlineEdges, validateManifest };
+//# sourceMappingURL=chunk-J3N3WRK7.mjs.map
+//# sourceMappingURL=chunk-J3N3WRK7.mjs.map